Can Good Governance Lower Financial Intermediation Costs?
Author:
Mariusz Jarmuzek
Search for other papers by Mariusz Jarmuzek in
Current site
Google Scholar
PubMedClose
and
Mr. Tonny Lybek
Search for other papers by Mr. Tonny Lybek in
Current site
Google Scholar
PubMedClose

Contributor Notes

Author’s E-Mail Address: mariusz_jarmuzek@hotmail.com and tlybek@imf.org
Publication Date:
11 Dec 2018
eISBN:
9781484385678
ISBN:
9781484385678
Language:
English
Keywords:
WP; intermediation cost; bank-dealership model; bank characteristic; banks capital; transaction size; capitalized bank; governance indicator; return on assets; returns to scale; funding cost
Download PDF (2.2 MB)

This paper argues that better governance practices can reduce the costs, risks and uncertainty of financial intermediation. Our sample covers high-, middle- and low-income countries before and after the global financial crisis (GFC). We find that net interest margins of banks are lower if various governance indicators are better. More cross-border lending also appears conducive to lower intermediation costs, while the level of capital market development is not significant. The GFC seems not to have had a strong impact except via credit risk. Finally, we estimate the size of potential gains from improved governance.

Abstract

This paper argues that better governance practices can reduce the costs, risks and uncertainty of financial intermediation. Our sample covers high-, middle- and low-income countries before and after the global financial crisis (GFC). We find that net interest margins of banks are lower if various governance indicators are better. More cross-border lending also appears conducive to lower intermediation costs, while the level of capital market development is not significant. The GFC seems not to have had a strong impact except via credit risk. Finally, we estimate the size of potential gains from improved governance.

I. Introduction

This paper examines whether financial intermediation costs of banks are influenced by the quality of governance across countries and over time. Following the global financial crisis (GFC) in 2008, banks’ net interest margins (net interest income to interest-earning assets: NIM) have received renewed attention (e.g., Burke and Garcia, 2017). They are typically employed as a good proxy for financial intermediation costs and efficiency, which are critical factors for supporting economic growth (Levine, 1997 and 2005). In principle, competition among banks should help reduce bank margins to levels constrained by expenses, risks and uncertainty.1 This paper aims at explaining cross-country variations in the NIM by differences in governance quality, while controlling for banking sector characteristics and macroeconomic indicators.2

The paper intends to fill some gaps in the literature. First, it merges various strands of literature on determinants of financial intermediation costs with the literature on the role of governance. It focuses on financial intermediation costs (price effects) as opposed to financial deepening (volume effects). Second, many of the empirical studies cover the period until the early 2000s, whereas this study also covers the post-GFC period (structural break). Third, the set of countries is large and includes advanced, emerging and developing economies. Fourth, compared to other studies, a broader set of governance indicators is examined. Fifth, various datasets are merged that shed light on features associated with international financial interconnectedness and capital market financing. Sixth, simulations quantifying potential gains of good governance in terms of financial intermediation costs are conducted. Finally, these findings should, in principle, help guide theoretical governance models that suggest opposing effects (see, for instance, Jappelli et al., 2005).

Our prior is that costs, risks and uncertainty caused by poor governance will ultimately be passed on to the customers and will hamper efficient financial intermediation and thus sustainable economic growth. Governance is a complex concept (e.g., Bevir, 2013). In this paper, good governance refers to legislation, formal and informal rules governing behaviors of society, institutions, organizations, firms, individuals, as well as networks and markets supporting the achievement of the official objective(s) of these entities. Such practices thus alleviate market-, public- and private-sector failures by lowering costs, risks, and uncertainty.3 Our focus is on governance in a broad context, like the effectiveness and efficiency of the legislative, executive and judiciary branches of government in order to improve the welfare and wealth of society.4 These factors obviously affect the behavior of individuals, non-financial corporations and banks, albeit somewhat distinct from the more specific G20/OECD (2015) Principles of Corporate Governance and the BIS (2015) Corporate Governance Principles for Banks. Corruption is often associated with, and is an important element of, poor governance (Shleifer and Vishny, 1993). However, poor governance can exist without corruption.5 Public governance practices—like property rights, the rule of law, efficient institutions, predictable contract enforcement, etc.—spill-over to private sector behaviors. For instance, good public governance practices are conducive to a healthier credit culture. Both lenders and borrowers are then driven by enlightened self-interest as the scope for pure self-interest (i.e., opportunistically exploiting information asymmetries and poor contract enforcement) diminishes.

In our investigation, we merge various strands of literature on determinants of financial intermediation costs with the literature on governance. Both theoretical and empirical studies suggest that elements of better governance—for instance effective institutions and less corruption—are positively correlated with higher per capita income (e.g., Mauro, 1995; Tanzi and Dawoodi, 2001), more efficient government spending, and greater financial deepening (e.g., Demirgüç-Kunt and Huizinga, 1999; Djankov et al., 2007; and Al Masry and Pedroni, 2016). However, instead of focusing on financial deepening (the quantity effect) we focus on financial intermediation costs (the price effect). Financial deepening likely reflects accumulated past price effects. In identifying factors shaping financial intermediation costs, we broadly follow the literature on determinants of NIM and efficiency of banks,6 which hinges on the bank-dealership model developed by Ho and Saunders (1981). Our empirical approach largely follows Demirgüç-Kunt and Huizinga (1999) and Demirgüç-Kunt et al. (2004).

We test if better governance practices do reduce NIM, and whether the impact changed after the GFC. Most similar cross-country empirical studies precede the GFC and cover smaller samples. A notable exception is the study by Poghosyan (2013), but it focuses on low-income countries and covers until 2010; before the GFC ended. Our database covers 100 countries spanning the period 1996 to 2015, although it is unbalanced across country and time dimensions. In line with the benchmark theoretical model, we find that countries with higher operating costs, risk aversion, and credit risk tend to have higher margins, while those with higher transaction size, i.e. more financial deepening, have lower margins. Consistent with the literature, we also find that various aspects of good governance significantly reduce NIM. The impact of the GFC on NIM appears to have worked mainly through credit risk. It has not led to a structural break in the relationship between governance and NIM.

Our investigation also sheds light on features associated with international financial interconnectedness and capital market development. While we observe that more financial openness seems to reduce NIM of banks, likely due to more competition, the results for the size of capital market financing are not significant. The latter finding is possibly because although good governance is expected to be conducive to capital market financing, it may also facilitate efficient bank intermediation.

Finally, we estimate potential gains from better governance and institutions for various country groups in order to draw policy conclusions. If governance improves to the top 10 percent threshold, the NIM would decline on average by about 1 percentage point for low-income countries and by about 0.3 percentage point for high-income countries. This difference is expected because high-income countries typically already have good governance, which reduces the scope for further gains. The “gains” in percent of GDP, however, are almost inverse, due to the greater size of bank intermediation (financial deepening) in high-income countries.

The rest of the paper has three main parts. Section II provides a basic conceptual framework. Section III presents the empirical analysis. It describes the data, empirical approach and results. Section IV provides some simple illustrative estimates of the costs of poor governance. Section V offers a brief conclusion.

II. Conceptual Framework: Net Interest Margins and Governance

The bank-dealership model (Ho and Saunders, 1981) considers banks to be risk averse, while intermediating deposits and loans. Credit institutions aim to maximize the expected net present value for their owners, while giving due consideration to risks and uncertainties. Their core activity is accepting deposits and extending loans to the general public. The bank maximizes its expected net present value by adjusting its spread on deposits and loans compared to the risk-free interest rate, as compensation for risks and uncertainty associated with mismatches between deposits and loans.

The model makes four general predictions, focussing on banking sector characteristics. First, banks that are able to exercise monopoly power can demand a greater spread compared to those operating in a more competitive market structure. Second, banks that have a higher risk aversion are likely to be better capitalized and are perceived as more solvent. In principle, this should reduce their funding costs and, thus, increase their margins. Third, given credit and market risks, the average transaction size should exert some impact on the margin due to returns to scale. Fourth, the more volatile interest rates, the greater the market risk and, therefore, the higher is the risk premium required by banks. These results have been broadly confirmed by empirical evidence, including by Ho and Saunders (1981) for US banks and by Saunders and Schumacher (2000) for banks in seven OECD countries.

A. Banking Sector Factors

The bank-dealership framework has been extended and operationalized. For example, Allen (1988) introduced heterogenous loans and Angbazo (1997) confirmed that the greater the uncertainty about granted loans, the greater the margin. Credit risk is often measured by the non-performing loan (NPL) ratio, provisions, or the provision coverage ratio. Maudos and Fernandez de Guevara (2004) incorporated several factors, including operating costs. The higher these costs, which may reflect inefficiencies related to poor governance, the higher the margin. The empirical evidence by Angbazo (1997) and Maudos and Fernandez de Guevara (2004) also suggested a negative relationship between the transactions size and margins, due to the economies of scale. Transaction size is often measured as bank credit to the private sector.

Different proxies are used to capture competition and risk aversion. Increased competition (less concentration) tends to lower margins, unless concentration is achieved through higher efficiency (Kasman and Corvallo, 2014). Often either the Herfindahl-Hirschman index or the share of the assets, deposits or loans of the five or three largest banks is used to measure concentration.7 Risk aversion is commonly proxied by the capital adequacy ratio and is frequently found to be positively associated with the NIM. Better capitalized banks are in principle less risky, which in turn should reduce their funding costs and consequently improve their interest margins. However, more capital is costly, as it reduces the leverage effect and may thus require a higher margin. Furthermore, a high capital ratio could incentivize banks to accept bigger risks.

B. Governance Factors

The literature has examined the link between various aspects of governance and bank intermediation costs. For instance, Jappelli et al. (2005) develop a theoretical model predicting that an improvement in contract enforcement increases aggregate lending by opening the credit market to borrowers with little collateral. Improvements in institutional quality increase the value of collateral, hence reducing the cost and risk of financial intermediation. There is also an opposing effect. Better governance and lower margins may reduce credit rationing, which could result in higher debt levels, including to less qualified borrowers. If banks have market power, they can increase the lending rate to account for this. But if there is strong competition, the model shows that the system could be more prone to instability.8

Empirical studies have confirmed that better governance, institutions and credit culture tend to lower financial intermediation costs and enhance financial deepening This has been documented by Demirgüç-Kunt and Huizinga (1999), Demirgüç-Kunt et al. (2004), and Poghosyan (2013) for different groups of countries. The governance proxies typically capture property rights, rule of law, legal system, regulatory quality, effectiveness of the executive branch,9 and efficiency of the judiciary. Corruption is more complicated, as it can affect all the steps in the intermediation process. Generally, it tends to increase intermediation costs,10 constrain efficient lending,11 and increase credit risk (e.g., Goel and Hasan, 2011).12 Some models, however, like Weill’s (2011), which uses “grease-the-wheels” argument, predicts that corruption in some cases can boost lending, particularly if banks are very risk averse. Relying on patronage and informal enforcement mechanisms that bypass the conventional ones are likely second-best solutions in the absence of good governance.

C. Macroeconomic Factors

Many past empirical papers have found that higher inflation is associated with higher interest spreads, but the results have been more mixed in newer studies on high-income countries. Inflation is usually associated with either stronger growth, hence higher credit demand, or monetary instability that contribute to higher risk and uncertainty. Moreover, inflation can erode banks capital, which also necessitates higher margins. On the other hand, inflation deflates clients’ debt and can facilitate their debt service. The former effects seem to dominate in earlier studies (e.g., Demirgüç-Kunt and Huizinga, 1999) and in low-income countries (Poghosyan, 2013), typically characterized by periods with higher inflation.

The empirical findings of the impact of real GDP growth on interest margins are even more mixed. A priori, one would expect that the NIM would increase during a boom, as demand for credit rises. However, during the early phase of a boom the risk-adjusted NIM may effectively decrease but the nominal NIM could remain rigid. During a recession, the intermediation cost may not adjust downwards in order to compensate for unexpected provisions and lower credit demand. The net impact will depend on whether quantity outpaces the price effects.

D. Empirical Methodologies

The empirical literature has essentially taken two approaches to examining bank intermediation costs. Ho and Saunders (1981) and Saunders and Schumacher (2000) pursue a two-stage approach when estimating bank interest margins. The initial stage involves isolating the so-called “pure margin” by introducing the variables that are not explicitly captured by the theoretical model. This is followed by a second stage that analyzes the relationship between the pure margin and the variables postulated by the theoretical model. This approach thus separates the effects on the margin for which macroeconomic policy is responsible and components for which the market structure is responsible. However, it requires relatively long timeseries.

The other approach is based on a single-stage methodology. McShane and Sharpe (1985) as well as Angbazo (1997) combine into one equation the variables suggested by the theoretical models with the variables that were not incorporated into modelling the pure margin in the two-stage approach. The main advantage is that this model does not require long timeseries. Applications of this approach include, among others, the work by Maudos and Fernandez de Guevara (2004), Demirgüç-Kunt et al. (2004), and Poghosyan (2013).

III. Explaining Bank Intermediation Costs

A. The Data

The dataset we use covers 100 countries spanning the period from 1996 to 2015. It is highly unbalanced across countries and time. During the earlier years, data are not available for many of the specifications.13 The NIM and other aggregated banking industry variables are mainly from Beck et al. (2017) based on Bankscope. The macro data are primarily from International Financial Statistics, IMF. The governance indicators are from Doing Business, World Bank, the Worldwide Governance Indicators based on Kaufmann, Kraay and Mastruzzi (2010), and ethics of private firms from the Global Competitiveness Index, World Economic Forum.14 The countries included in the sample are listed in Appendix I, the data are summarized in Appendix II and the correlation coefficients are tabled in Appendix III.

Financial Intermediation Costs—The Net Interest Margin

Financial intermediation costs of banks can be measured in different ways that capture various aspects of the governance quality and credit culture. Interest spreads only capture the difference between lending and deposit rates, but not changes in the composition of the balance sheet nor remuneration of required reserves. Return on assets (ROA) does not reflect the operating costs of doing business, but capture compensation for risks, uncertainty and taxation to the extent that ROA before extraordinary items and taxation is used. We, therefore, follow the predominant practice by using the net interest margin (NIM)—i.e., net interest income in percent of interest earning assets—as it captures compensation for costs, risks and uncertainty.

There is a significant variation in bank intermediation costs across countries, regions and income groups. High-income countries in Asia, Europe and North America are characterized by the lowest margins and low-income countries in Sub-Saharan Africa, Latin America and the Caribbean have the highest margins. NIM for high-income countries are on average below 3 percent and do not exceed 10 percent. This is in sharp contrast with low-income countries, which have twice as high NIM, with many of them having NIM exceeding 15 percent (Figure 1).

Figure 1.
Figure 1.

Net Interest Margins in Different Country Groups, 1996 – 2015

Citation: IMF Working Papers 2018, 279; 10.5089/9781484385678.001.A001

Source: Beck et al. (2017).Note: The median is indicated by a line subdividing the interquartile range (the box) with whiskers indicating the lower and upper adjacent values.

The NIM of banks across time do not reveal a strong pattern (Figure 2). The average bank intermediation costs do not visibly change over time spanning the period from 1996 to 2015. However, there seems to be a modest downward trend from 2013. Interestingly, the GFC did not bring any significant change in the NIM trend. The policy response to the GFC in many countries involved substantial monetary policy loosening, typically in the form of lowering interest rates but also nonstandard monetary policy actions that often involved some form of quantitative easing. These actions affected both deposit and lending rates. While the funding costs was limited by the zero lower bound, competition for lending to viable projects may have further driven down lending rates, hence causing some compression of the NIM, especially in European countries from 2013. Nevertheless, net interest income in the Euro Area declined only marginally according to Burke and Garcia (2017) and the NIM remained broadly stable over the financial cycle according to Detragiache et al. (2018).

Figure 2.
Figure 2.

Net Interest Margins over Time, 1996 – 2015

Citation: IMF Working Papers 2018, 279; 10.5089/9781484385678.001.A001

Source: Beck et al. (2017).Note: The median is indicated by a line subdividing the interquartile range (the box) with whiskers indicating the lower and upper adjacent values.

B. The Econometric Analysis

The econometric analysis covers four main dimensions. The initial dimension was estimated simply by pooling the data. The pooled OLS estimator was used, and then extended by the inclusion of time effects to capture global business and financial cycles. Second, drawing on the literature applying panel data techniques to examine the role of institutions, we follow Plumber et al. (2005, 2007) as well as Bell and Jones (2015) and employ the random effect model. This allows accounting for dummy variables capturing certain institutional features as well as the fact that governance indicators typically are very slow-moving. We examine the relevance of the random effect by the Breusch-Pagan (1979) test. Third, we address endogeneity by using instrument variable estimations. Fourth, since cross-sectional dependence can be an important feature in macro-panels (Chudik and Pesaran, 2015), we use the Pesaran (2006) test and subsequently proceed with the regressions using standard errors to account for this phenomenon following Driscoll and Kraay (1998).

Our analysis uses a single-stage methodology. Specifically, following the existing literature, we estimate variants of the following equation:

N I M i t = α 0 + Σ k = 1 5 α 1 , k B S V k , i t + Σ k = 1 2 α 2 , k M A C k , i t + Σ k = 1 M α 3 , k G O V k , i t + Σ k = 1 N α 4 , k O T H k , i t + ϵ i t

where NIM denotes banks’ NIM in country i at time t, which are explained by banking sector variables (BSV) that comprise measures of competition, operating cost, transaction size, risk aversion, and credit risk; macroeconomic indicators (MAC) like inflation and real economic growth; governance indicators (GOV) that include proxies for the rule of law, regulatory quality, recovery of assets, perceived corruption, ethics of private firms, etc.; and other variables (OTH) that are considered in various extensions.

Basic Specification

A simple regression analysis was conducted, encompassing a basic set of bank characteristics and macroeconomic variables (Table 1). The ambiguity of concentration is confirmed by the data. In the basic specifications we find that our measure for more concentration is significantly correlated with a smaller NIM. In the extended models, the coefficient is also negative, but is only significant in a few of the specifications. In principle, higher concentration should mean more market power, less competition and hence higher NIM. However, there may be returns to scale and efficiency gains. If the latter factor dominates, our measure of larger concentration would imply lower NIM. Given our measure of competition—share of assets of the three largest banks—the impact of concentration is broadly consistent with theory and results of previous studies.15

Table 1.

Basic Specification

article image
Sources: IMF staff estimates. Note: Figures in parentheses indicate standard errors. An *, **, and *** mean significance at, respectively, 10, 5 and 1 percent levels.

The other banking sector variables suggested by the dealership model turn out to be statistically significant. Overhead cost is found to be positively associated with NIM, reflecting that banks tend to pass these costs on to their clients, in line with the results obtained by Williams (2007). The transaction size is negatively associated with NIM, pointing to the importance of increasing returns to scale in banking, consistent with previous findings. Risk aversion is positively associated with NIM, reflecting that banks that are better capitalized should in turn have lower funding costs and consequently higher NIM, which is in line with earlier results. Credit risk is positively associated with NIM, confirming that banks charge additional risk premium to compensate for ex post credit risk, which is also consistent with previous findings.16

Like most other more recent studies, the macroeconomic variables, with some exceptions of inflation, are not robustly statistically significant. The impact of real GDP growth17 and in many cases CPI inflation are not statistically significant, likely due to the counteracting effects discussed in Section II. The exception is the pooled basic model, where inflation is positively associated with higher NIM.

Extended Models

The basic model is extended with various governance indicators to test their relevance on the NIM (Table 2). Like previous studies, we found overwhelming evidence that better governance reduces the NIM. The governance proxies spanned from comprehensive indices on the rule of law to indicators focusing narrowly on the insolvency framework, contract enforcement, and government effectiveness. This can for instance be interpreted that improvements in governance tend to increase the value of collateral for bank loans and consequently reduce the cost of bank intermediation, which is consistent with the evidence by Demirguc-Kunt and Huizinga (1999), Demirguc-Kunt et al. (2004), and Poghosyan (2013).18 Given that poor governance may lead to higher corruption, we also test for the importance of this variable and find supporting evidence that more corruption tends to increase the NIM, consistent with Poghosyan (2013). With a view to also gauge governance quality in the private sector, we also test ethical behavior of private firms, where higher standards also are associated with lower NIM.19

Table 2.

Models Extended with Selected Governance Variables

article image
Sources: IMF staff estimates. Note: The random effects general least squares (GLS) estimator was used. Figures in parentheses indicate standard errors. An *, **, and *** mean significance at, respectively, 10, 5 and 1 percent levels.

Greater international interconnectedness appears to lower the NIM (Table 3). In principle, more openness would limit how much the NIM can increase before clients will solicit funding from abroad. Higher levels of both international debt and cross-border bank loans, which are used to indicate greater international interconnectedness, tend to reduce the NIM, while controlling for financial deepening and macroeconomic variables.20 Countries with more financial deepening are likely also to be more interconnected.

Table 3.

Extended Models with Interconnectedness

article image
Sources: IMF staff estimates. Note: The random effects general least squares (GLS) estimator was used. Figures in parentheses indicate standard errors. An *, **, and *** mean significance at, respectively, 10, 5 and 1 percent levels.

Deeper private capital markets do not appear to significantly reduce the NIM (Table 4). We used stock market capitalization, traded stocks in percent of GDP and private bond issuances as proxies for capital market development. A previously mentioned, well-developed capital markets depend on a solid legal system with good governance practices. These conditions are also advantageous for banking and may thus outweigh much of the competition from the capital market.21

Table 4.

Extended Models with Capital Market Variables

article image
Sources: IMF staff estimates. Note: The random effects general least squares (GLS) estimator was used. Figures in parentheses indicate standard errors. An *, **, and *** mean significance at, respectively, 10, 5 and 1 percent levels.

The Global Financial Crisis

The potential impact of the GFC on the NIM is explored (Table 5). This extension is meant to examine: (i) if there was any significant change in the NIM before and after the GFC; and (ii) whether the relationship between explanatory variables and the NIM changed with the GFC. A GFC dummy (1 in 2007 and afterwards) is thus included in the regressions. There is only very limited evidence that the GFC tended to reduce the NIM, with the GFC dummy being significant in less than half the specifications.

Table 5.

Extended Models with Impact of Global Financial Crisis

article image
Sources: IMF staff estimates. Note: The random effects general least squares (GLS) estimator was used. Figures in parentheses indicate standard errors. An *, **, and *** mean significance at, respectively, 10, 5 and 1 percent levels.

The interaction terms suggest that the GFC affected the NIM mainly via credit risk. Interactive variables are constructed by multiplying the GFC dummy by the respective explanatory variable. This is to assess whether the GFC worked indirectly via the other variables. The results generally indicate no statistically significant role of these interaction terms, with the notable exception of credit risk, which is overpowered by the interaction term. It seems that credit risk is much more strongly associated with the NIM after the GFC. This in line with the findings by Apergis and Cooray (2018).22 They argue that credit risk has become more appropriately reflected after the GFC, which suggests that it was underpriced prior to the GFC. After the crisis, more costly screening, monitoring, and higher capital requirements have been introduced (Mishkin and Eakins, 2012). The pertinent interaction terms, however, are not significant.

C. Robustness Checks

Additional governance indicators showed similar results (Table 10). These indicators included property right protection, judicial independence and impartiality of courts. Although the coverage of these indicators is not as good as the ones employed in the main specifications, they generally show similar results, pointing to the importance of governance in setting net interest margins by banks.23

Instrument variables were used to account for endogeneity (Table 11). Given the fact that governance indicators may potentially be endogenous, we draw on the political economy literature and employ as instruments voice and accountability as well as political stability indicators. They are considered to shape political power and institutions, which in turn shape economic institutions. The latter typically is captured by the rule of law, regulatory and insolvency framework and contract enforcement, as discussed by, among others, Acemoğlu (2005) and Acemoğlu and Robinson (2016).24 We also include lagged explanatory variables as instruments.

Adjustments are made for cross-sectional dependence (Table 12), which can be an important issue in macro-panels (Chudik and Pesaran, 2015). To test for the presence of cross-sectional dependence, we follow Pesaran (2006) and subsequently proceed with the regressions using standard errors accounting for this phenomenon following Driscoll and Kraay (1998).

A probit model is used to account for possible non-linearity in the relationship between net interest margins and candidate explanatory variables (Table 13). Certain mechanisms may start operating only at higher level of margins. A probit model is thus applied with the cut-off point for the dependent variable set at the 75th percentile of the distribution. The results did not change significantly compared to the basic regressions.

Panel estimations with lagged dependent variable using standard techniques can be challenging (Table 14). Judson and Owen (1999) have shown that the standard estimator performs better than other methods, such as the generalized method of moments estimator, when N is moderately large and T is relatively small, as in this paper. The results with the lagged dependent variable tend to reduce the coefficients somewhat, but do not materially change the across the various specifications.25

Finally, the sensitivity to outliers is investigated (Table 15). Following Blanchard and Leigh (2013), we re-estimate the specifications using robust regressions, which down-weighs observations with larger absolute residuals using iterative weighted least squares (Andersen, 2008). This methodology results in broadly similar results across the various specifications.26

IV. Policy Simulations: The Cost of Poor Governance

The above empirical findings offer an opportunity to gauge the financial intermediation costs caused by poor governance. The NIM of the seven models shown in Table 2 are estimated to the extent data are available.27 The simple average fitted NIM of these models is calculated using the actual values for each year. Then the average fitted NIM is calculated assuming that the respective governance variable instead is at the top 10 percent threshold. The difference between these two estimations, all other things equal, shows how much lower the NIM could have been due to improved governance. The NIM differences are then multiplied on bank credit to the private sector in percent of GDP to evaluate the potential annual gains, or savings, in percent of GDP. In other words, the same level of intermediation could have been achieved with this amount of fewer resources, had the governance been at the top 10 percent level.

Policy simulations suggest that the potential gains from better governance can be sizeable (Table 6). Low-income countries have more room to improve their governance and could thus reduce their NIM by on average about 1 percentage point and high-income countries on average by about 0.3 percentage point. (High-income countries at or above the top 10 percent governance threshold were not included.) The gains in percent of GDP are almost inverse, due to the greater financial deepening in higher income countries. The annual savings may be sizable accumulated over time. This could explain why it seems to be easier to find positive association between governance and financial deepening than with financial intermediation costs. Furthermore, the estimated annual gains may not fully capture other beneficial side-effects, like lengthening of maturities, less required collateral due to better governance, etc.28

Table 6.

Estimated Potential Savings from Improved Governance

article image
Source: Authors’ calculations. Note: Calculated as the simple average of annual estimated savings, whenever data are available; i.e. for some countries only one or a few years were included. The savings are calculated as the difference between the fit of the actual values and the fit if the pertinent governance variable had been at the top 10 percent threshold. The simple average of up to seven models shown in Table 2 were used to the extent data were available, i.e. some of the annual observations reflect the average of fewer than the seven models.

Simple averages of annual savings, i.e., the lower net interest margin due to the assumed higher governance quality was applied on bank credit to the private sector in percent of GDP. For some countries (Belgium, France, Ireland, Portugal and Spain) at least one observation could not be included due to lack of data on bank credit, but they were still included in the calculated of the net interest margin (second column in this Table).

Simple average of lower net interest margins if the pertinent governance variable – at a minimum in one of the seven models in Table 2 – was increased to the top 10 percent threshold.

Number of countries where data were available for at a minimum one year in – at a minimum one of the seven models in Table 2 – during the 1996–2015 period.

Some high-income country observations were not included in the calculation of savings. Mainly because their governance indicators were already at the top 10 percent or because the use of the average of the seven models resulted in negative savings in some years. The following countries were not included: Denmark, Finland, Netherlands, Norway, and Singapore. The following countries had one or more of the annual calculations excluded: Australia, Austria, Canada, France Germany, Sweden, Switzerland, and the United Kingdom.

Improving the more general governance indicators to the specified threshold typically seem to have the biggest impact. Enhancing the rule of law, boosting the effectiveness of government and reducing corruption appear to have the largest impact on the NIM. Interestingly, improving the ethics of private firms seem to have an even bigger impact, likely because it points to a better credit culture. The number of observations, however, is much smaller for ethics of private firms. Obviously, priority should be given to eliminate the most harmful practices. Nevertheless, the consequences for the whole economy must be considered before introducing specific measures primarily focused on lowering financial intermediation costs. Moreover, low-hanging fruits should be given priority, i.e. prioritize the path of least resistance. The visible benefits of such an approach could pave the way for additional reforms. Finally, although new technology poses its own governance challenges, it also offers new opportunities that enable lower income countries to leapfrog, as monitoring and information gathering become much more affordable.

These results are only illustrative and must be interpreted with caution. Most of the governance indicators were not available in the earlier part of the analyzed period and for some countries only a few years were included. Some bank characteristics may not be entirely independent of the governance indicators. While improving the governance quality to the top 10 percent level may seem exceptionally ambitious, there could be improvements in several governance indicators at the same time. Moreover, there could be other positive externalities, including on sustainable real growth. In short, the simulations lend credence to the intuition that better perceived governance practices should reduce costs, risks and uncertainty.

V. Conclusion

Our empirical analysis provides evidence that good governance practices are associated with lower financial intermediation costs. If governance practices were to improve to the top 10 percent threshold, the NIM of low-income countries would decline on average by about 1 percentage point and high-income countries by about 0.3 percentage point, as the latter’s governance practices already are better. The “gains” in percent of GDP, however, are almost inverse, due to the greater financial deepening in higher income countries. In many countries, it could amount to about ⅓ percent of GDP per year. Accumulated over time, such savings could become sizable.

The paper reconfirms that various bank characteristics significantly influence bank intermediation costs. Countries with higher operating costs, risk aversion as well as credit risk tend to have higher margins, while those with higher transaction size have lower margins, hinting at increasing returns to scale.

The impact of the GFC on the NIM appears to have worked mainly through credit risk. Credit risk is more strongly associated with the NIM after the GFC compared to the period prior to the crisis. This may stem from the fact that some of the lending prior to the GFC proved to be too lax and unsustainable.29

Further research is needed to fully explore why improvements in governance are not pursued more proactively to realize the substantial gains. This is likely to be a typical collective action problem, in which the benefits are diffuse and widespread, hence difficult to internalize. At the same time, those benefitting from poor governance are often well-organized, influential and are able to habitually pass the costs on to their clients and tax payers.

References

  • Acemoğlu, Daron; and James A. Robinson, 2016, “Paths to Inclusive Political Institutions,” https://www.researchgate.net/publication/308340491_Paths_to_Inclusive_Political_Institutions.

    • Search Google Scholar
    • Export Citation

  • Acemoğlu, Daron; and James A. Robinson, 2012, Why Nations Fail: The origins of Power, Prosperity, and Poverty, Crown Publishers, New York.

    • Search Google Scholar
    • Export Citation

  • Acemoğlu, Daron; Simon Johnson; and James A. Robinson, 2008, “The Role of Institutions in Growth and Development, Working Paper No. 10, Commission on Growth and Development, World Bank.

    • Search Google Scholar
    • Export Citation

  • Acemoğlu, Daron, 2005, “Constitutions, Politics, and Economics: A Review Essay on Persson and Tabellini’s the Economic Effects of Constitutions,” Journal of Economic Literature, No. 43, pp. 1025–1048.

    • Search Google Scholar
    • Export Citation

  • Acemoğlu, Daron; Simon Johnson; and James A. Robinson, 2001, “The Colonial Origins of Comparative Development: An Empirical Investigation,”, The American Economic Review, Vol. 91. No. 5, December, pp. 1369–1401.

    • Search Google Scholar
    • Export Citation

  • Akins, Brian; Yiwei Dou; and Jeffrey Ng, 2017, “Corruption in bank lending: The role of timely loan loss recognition,” Journal of Accounting and Economics, Vol. 63, pp. 454–478.

    • Search Google Scholar
    • Export Citation

  • Akitoby, Bernadin; and Thomas Stratmann, 2010, “The value of institutions for financial markets: evidence from emerging markets,” Review of World Economics, Vol. 146, Issue 4, pp. 781–797.

    • Search Google Scholar
    • Export Citation

  • Allen, Linda, 1988, “The Determinants of bank Interest Margins: A Note,” Journal of Financial and Quantitative Analysis, Vol. 23, No. 2, June, pp. 231–235.

    • Search Google Scholar
    • Export Citation

  • Al Masri, Diala; and Peter Pedroni, 2016, “Nonlinearities in Financial Development; The Good, the Bad and the Beautiful,” Working Paper Sponsored by The Casey Lecture Fund for Economics.

    • Search Google Scholar
    • Export Citation

  • Andersen, Robert, 2008, Modern Methods for Robust Regression (Thousand Oaks, California: SAGE Publications).

  • Angbazo, Lazarus, 1997. “Commercial bank net interest margins, default risk, interest-rate risk and off-balance sheet banking,” Journal of Banking & Finance 21, pp. 55–87.

    • Search Google Scholar
    • Export Citation

  • Apergis, Nicholas; and Arusha Cooray, 2018, “The Behavior of Interest Rate Spreads Prior to and After the Financial Crisis: Evidence Across OECD Countries,” The Manchester School, Vol. 86, No. 5, September, pp. 559–589.

    • Search Google Scholar
    • Export Citation

  • Arrow, Kenneth, J., 1974, The Limits of Organization, Norton and Company Inc., New York.

  • Bae, Kee-Hong; and Vidhan K. Goyal, 2009, “Creditor Rights, Enforcement, and Bank Loans,” The Journal of Finance, Vol. 64, issue 2, pp. 823–860.

    • Search Google Scholar
    • Export Citation

  • Barth, James R.; Chen Lin; Yue Ma; Jesús Seade; Frank M. Song, 2013, “Do bank regulation, supervision and monitoring, enhance or impede bank efficiency?” Journal of Finance & Banking, Vol. 36, pp. 2879–2892.

    • Search Google Scholar
    • Export Citation

  • Barth, James R.; Chen Lin; Ping Lin; and Fran M. Song, 2009, “Corruption in bank lending to firms: Cross-country micro evidence on the beneficial role of competition and information sharing,” Journal of Financial Economics, Vol. 91, Elsevier, pp. 361–388.

    • Search Google Scholar
    • Export Citation

  • Beck, Thorsten; Aslı Demirgüç-Kunt; Ross Levine; Martin Cihak; and Erik H. B. Feyen, 2017, “Financial Development and Structure Dataset,” World Bank, Washington DC. http://www.worldbank.org/en/publication/gfdr/data/financial-structure-database

    • Search Google Scholar
    • Export Citation

  • Beck, Thorsten; Aslı Demirgüç-Kunt; and Ross Levine, 2006, “Bank supervision and corruption in lending,” Journal of Monetary Economics, Vol. 53, pp. 2121–2163.

    • Search Google Scholar
    • Export Citation

  • Bell, Andrew; and Kelvyn Jones, 2015, “Explaining fixed effects: random effects modelling of time-series cross-sectional and panel data,” Political Science Research and Methods, Vol. 3, No. 1, pp. 133–153.

    • Search Google Scholar
    • Export Citation

  • Bevir, Mark, 2013, A Theory of Governance, University of California Press Ltd., Berkeley and Los Angeles.

  • BIS, 2015, Guidelines: Corporate Governance Principles for Banks, Basel Committee on Banking Supervision, Bank for International Settlements, Basel.

    • Search Google Scholar
    • Export Citation

  • Blanchard, O.; and D. Leigh, 2013, “Growth Forecast Errors and Fiscal Multipliers,” American Economic Review: Papers and Proceedings, Vol. 103 No. 3, pp. 117–120.

    • Search Google Scholar
    • Export Citation

  • Boyd, John; and Bruce Smith, 1996, “The Coevolution of the Real and Financial Sectors in the Growth Process,” The World Bank Economic Review, Vol. 10, No. 2, pp. 371–396.

    • Search Google Scholar
    • Export Citation

  • Breusch, T. S.; and A. R. Pagan, 1979, “A Simple Test for Heteroscedasticity and Random Coefficient Variation, Econometrica, Vol. 47 No. 5, September, pp. 1287–1294.

    • Search Google Scholar
    • Export Citation

  • Bruno, G., 2005, “Approximating the Bias of the LSDV Estimator for Dynamic Unbalanced Panel Data Models,” Economics Letters, Vol. 87, pp. 361–366.

    • Search Google Scholar
    • Export Citation

  • Bun, M. J. G.; and J. F. Kiviet, 2003, “On the diminishing returns of higher order terms in asymptotic expansions of bias,” Economics Letters, Vol. 79, pp. 145–152.

    • Search Google Scholar
    • Export Citation

  • Burke, Javier Villar; and Cristina T. Plata Garcia, 2017, “Interest margins and efficiency: The impact of the crisis across EU countries,” BBVA Research Watch, June 28.

    • Search Google Scholar
    • Export Citation

  • Chudik, Alexander; and M. Hashem Pesaran, 2015, “Common correlated effects estimation of heterogenous dynamic panel data models with weakly exogenous regressors, Journal of Econometrics, Vol 188, pp.

    • Search Google Scholar
    • Export Citation

  • Chortareas, Georgios E; Claudia Girardone; and Alexia Ventouri, 2012, “Bank supervision, regulation and efficiency: Evidence from the European Union,” Journal of Financial Stability, Vol. 8, pp. 292–302.

    • Search Google Scholar
    • Export Citation

  • Claessens, Stijn; and Luc Laeven, 2004, “What Drives Bank Competition? Some International Evidence,” Journal of Money, Credit and Banking, Vol. 36, No. 3, Part 2, June, pp. 563–583.

    • Search Google Scholar
    • Export Citation

  • Claeys, Sophie; and Rudi Vander Vennet, 2008, “Determinants of bank interest margins in Central and Eastern Europe: A comparison with the West,” Economic Systems, Vol. 32, pp. 197–216.

    • Search Google Scholar
    • Export Citation

  • Coase, Ronald H., 1960, “The Problem of Social Cost,” Journal of Law and Economics, Vol. III, October, pp. 1–44.

  • Coleman, James S., 1988, “Social Capital in the Creation of Human Capital,” American Journal of Sociology, Vol. 94, Supplement: Organization and Institutions: Sociological and Economic Approaches to the Analysis of Social Structure, The University of Chicago Press, pp. 95–120.

    • Search Google Scholar
    • Export Citation

  • Demirgüç-Kunt, Asli; Luc Laeven; and Ross Levine, 2004, “Market Structure, Institutions, and the Cost of Financial Intermediation,” Journal of Money, Credit and Banking, Vol. 36, No. 3, Part 2, pp. 593–622.

    • Search Google Scholar
    • Export Citation

  • Demirgüç-Kunt, Asli; and Harry Huizinga, 1999, “Determinants of Commercial Bank Interest Margins and Profitability: Some International Evidence,” The World Bank Economic Review, Vol. 13, No. 3, pp. 379–408.

    • Search Google Scholar
    • Export Citation

  • Detragiache, Enrica; Thierry Tressel; and Rima Turk-Ariss, 2018, “Where Have All the Profits Gone? European Bank Profitability Over the Financial Cycle,” Working Paper WP/18/99, International Monetary Fund, Washington DC.

    • Search Google Scholar
    • Export Citation

  • Djankov, Simeon; Caralee McLiesh; and Andrei Shleifer, 2007, “Private Credit in 219 Countries,” Journal of Financial Economics, Vol. 84, Elsevier, pp. 299–329.

    • Search Google Scholar
    • Export Citation

  • Driscoll, John C.; and Aart C. Kraay, 1998, “Consistent Covariance Matrix Estimation with Spatially Dependent Panel Data, The Review of Economics and Statistics, pp. 549–560.

    • Search Google Scholar
    • Export Citation

  • Fisher, Irving, 1933, “The Debt-Deflation Theory of Great Depressions,” Econometrica, Vol.1, No. 4, October, pp. 337–357.

  • Fukuyama, Francis, 1995, Trust: The Social Virtues and the Creation of Prosperity, The Free Press, New York.

  • Fungáčová, Zuzana; and Tigran Poghosyan, 2011, “The Determinant of bank interest margins in Russia: Does Bank ownership matter?” Economic Systems, Vol. 35, Elsevier, pp. 481–495.

    • Search Google Scholar
    • Export Citation

  • Goel, Rejeev K.; and Iftekhar Hasan, 2011, “Economy-wide corruption and bad loans in banking: international evidence,” Applied Financial Economics, Vol. 21:7, pp. 455–461.

    • Search Google Scholar
    • Export Citation

  • Hall, R.; and C. Jones (1999). “Why do some countries produce so much more output per worker than others?”, Quarterly Journal of Economics, 114 (1), pp.83–116

    • Search Google Scholar
    • Export Citation

  • Ho, Thomas S. Y.; and Anthony Saunders, 1981, “The Determinants of Bank Interest Margins: Theory and Empirical Evidence,” The Journal of Financial and Quantitative Analysis, Vol. 16, No. 4, Proceedings of the 16th Annual Conference of the Western Finance Association, pp. 581–600.

    • Search Google Scholar
    • Export Citation

  • Houston, Joel F.; Chen Lin; and Yue Ma, 2011, “Media ownership, concentration and corruption in bank lending,” Journal of Financial Economics, Vol. 100, 326–350.

    • Search Google Scholar
    • Export Citation

  • IMF, 2018, Review of the 1997 Guidance Note on Governance—A proposed Framework for Enhanced Fund Engagement, April, International Monetary Fund, Washington DC.

    • Search Google Scholar
    • Export Citation

  • IMF, 2017, “Reforming the Judiciary: Learning from the Experience of Central, Eastern, and Southeastern Europe,” chapter 2 in Regional Economic Outlook, Europe, November, European Department, International Monetary Fund, Washington DC.

    • Search Google Scholar
    • Export Citation

  • Islam, Roumeen, 2007, “Economic Information and Finance: More information means more credit, fewer bad loans and less corruption,” Policy Research Working Paper, No. 4250, World Bank, Washington DC.

    • Search Google Scholar
    • Export Citation

  • Jappelli, Tullio; Marco Pagano; and Magda Bianco, 2005, “Courts and Banks: Effects of the Judicial Enforcement on Credit Markets,” Journal of Money, Credit and Banking, Vol. 37, No. 2, April, pp. 223–244.

    • Search Google Scholar
    • Export Citation

  • Jappelli, Tullio; and Marco Pagano, 2002, “Information sharing, lending and defaults: Crosscountry evidence, Journal of Banking & Finance, Vol. 26, pp. 2017–2045.

    • Search Google Scholar
    • Export Citation

  • Judson, Ruth A.; and Ann L. Owen, 1999, “Estimating dynamic panel data models: a guide for macroeconomists,” Economics Letters Vo. 65, pp. 9–15.

    • Search Google Scholar
    • Export Citation

  • Kasman, Adnan; and Oscar Carvallo, 2014, “Financial Stability, Competition and Efficiency in Latin American and Caribbean Banking,” Journal of Applied Economics, Vol. XVII, No. 2, November, pp. 301–324.

    • Search Google Scholar
    • Export Citation

  • Kasman, Adnan; Saadet Kasman; and Evrim Turgutlu, 2011, “Testing Profit and Structure Relationship in the European Banking Markets Using Efficiency Measures,” The Developing Economies, Vol. 49, No. 4, pp. 404–428.

    • Search Google Scholar
    • Export Citation

  • Kasman, Adnan, 2010, “Consolidation and Competition in the Banking Industries of the EU Member and Candidate Countries,” Emerging Markets and Finance and Trade, Vol. 46, No. 6, pp. 121–39.

    • Search Google Scholar
    • Export Citation

  • Kaufmann, Daniel; Aart Kraay; and Massimo Mastruzzi, 2010, “The Worldwide Governance Indicators: Methodological and Analytical Issues,” Policy Research Working Paper No. 5430, World Bank, Washington. The WGI database is updated regularly and is available on: http://info.worldbank.org/governance/wgi/#home

    • Search Google Scholar
    • Export Citation

  • Kiviet, J. F., 1999, “Expectation of expansions for estimators in a dynamic panel data model; some results for weakly exogenous regressors,” in Analysis of Panels and Limited Dependent Variable Models, edited by C. Hsiao, K. Lahiri, L.-F. Lee, and M. H. Pesaran, Cambridge University Press, pp. 199–225.

    • Search Google Scholar
    • Export Citation

  • Kiviet, J. F., 1995, “On bias, inconsistency, and efficiency of various estimators in dynamic panel data models,” Journal of Econometrics, Vol. 68, pp. 53–78.

    • Search Google Scholar
    • Export Citation

  • Knack, Stephen; and Philip Keefer, 1997, “Does Social Capital Have and Economic Payoff? A Cross-Country Investigation,” The Quarterly Journal of Economics, Vol. 112, No. 4, November, pp. 1251–1288.

    • Search Google Scholar
    • Export Citation

  • Knight, F. H., 1921, Risk, Uncertainty, and Profit, Boston, MA: Hart, Schaffner & Marx; Houghton Mifflin Company.

  • Levine, Ross, 2005, “Finance and Growth: Theory and Evidence,” chapter 12 in Handbook of Economic Growth, Volume 1A, edited by Philippe Aghion and Steven N. Durlauf, Elsevier B.V.

    • Search Google Scholar
    • Export Citation

  • Levine, Ross, 1997, “Financial Development and Economic Growth: Views and Agenda,” Journal of Economic Literature, Vol. 35, No. 2, June, pp. 688–726.

    • Search Google Scholar
    • Export Citation

  • Liebeg, David; and Markus S. Schwaiger, 2009, “What Drives the Interest Rate Margin Decline in EU Banking – The Case of Small Local Banks,” Kredit und Kapital, No. 4, pp. 509–538.

    • Search Google Scholar
    • Export Citation

  • Maudos, Joaquín; and Juan Fernández de Guevara, 2004, “Factors explaining the interest margin in the banking sectors of the European Union,” Journal of Banking and Finance, Vol. 28, pp. 2259–2281.

    • Search Google Scholar
    • Export Citation

  • Mauro, Paolo, 1995, “Corruption and Growth,” The Quarterly Journal of Economics, Vol. 110, No. 3, August, pp. 681–712.

  • McShane, R.W.; and I. G. Sharpe, 1985, “A Time Series/Cross Section Analysis of The Determinants of Australian Trading Bank Loan/Deposit Interest Margins: 1962–1981,” Journal of Banking and Finance, Vol. 9, North-Holland, pp. 115–136.

    • Search Google Scholar
    • Export Citation

  • Mishkin, Frederic S.; and Stanley G. Eakins, 2012, Financial Markets and Institutions, 7th edition, The Prentice Hall series in finance.

    • Search Google Scholar
    • Export Citation

  • North, Douglas C., 1991, “Institutions,” Journal of Economic Perspectives, Vol. 5, No. 1, Winter, pp. 97–112.

  • North, Douglas C., 1989, “Institutions and Economic Growth: An Historical Introduction,” World Development, Vol. 17, No. 9, Pergamon Press Plc., pp. 1319–1332.

    • Search Google Scholar
    • Export Citation

  • North, Douglas C., 1981, Structure and Change in Economic History, W.N. Norton & Co., New York.

  • OECD, 2015, G20/OECD Principles of Corporate Governance, OECD Publishing, Paris.

  • OECD, 2013, International Drivers of Corruption: A Tool for Analysis, OECD Publishing, Paris.

  • Padilla, Jorge A; and Marco Pagano, 1997, “Endogenous Communication Among Lenders and Entrepreneurial Incentives,” The Review of Financial Studies, Vol. 10, No. 1, pp. 205–236.

    • Search Google Scholar
    • Export Citation

  • Pesaran, M. Hashem, 2006, “Estimation and Inference in Large Heterogenous Panels with a Multifactor Error Structure,” Econometrica, Vol. 74, No. 4, July, pp. 967–1012.

    • Search Google Scholar
    • Export Citation

  • Plumper, Thomas; and Vera E. Troeger, 2007, “Efficient Estimation of Time-invariant and Rarely Changing Variables in Finite Sample Panel Analyses with Unit Fixed Effects,” Political Analysis, Vol. 15, No. 2, pp. 24–39.

    • Search Google Scholar
    • Export Citation

  • Plumper, Thomas; Vera Troeger; and Philip Manow, 2005, “Panel Data Analysis in Comparative Politics. Linking Methods to Theory,” European Journal of Political Research, Vol. 44, No. 2, pp. 327–354.

    • Search Google Scholar
    • Export Citation

  • Poghosyan, Tigran, 2013, “Financial Intermediation Costs in Low Income Countries: The role of Regulatory, Institutional, and Macroeconomic Factors,” Economic Systems, Vol 37, Elsevier B. V., pp. 92–110.

    • Search Google Scholar
    • Export Citation

  • Putterman, Louis, 2013, “Institutions, social capability, and economic growth, Economic Systems, Vol. 37, Elsevier B. V., pp. 345–353.

    • Search Google Scholar
    • Export Citation

  • Qian, Jun; and Philip E. Strahan, 2007, “How Laws and Institutions Shape Financial Contracts: The Case of Bank Loans,” The Journal of Finance, Vol. LXII, No. 6, pp. 2803–2834.

    • Search Google Scholar
    • Export Citation

  • Saunders, Anthony; and Liliana Schumacher, 2000, “The determinants of bank interest rate margins: an international study,” Journal of International Money and Finance, Vol. 19, pp. 813–832.

    • Search Google Scholar
    • Export Citation

  • Shleifer, Andrei; and Robert W. Vishny, 1993, “Corruption,” The Quarterly Journal of Economics, Vol. 108, No. 3, August, pp. 599–617.

    • Search Google Scholar
    • Export Citation

  • Staikouras, Christos; and Geoffrey E. Wood, 2004, “The Determinants of European Bank Profitability,” International Business & Economics Research Journal, Vol. 3, No. 6, pp. 57–68.

    • Search Google Scholar
    • Export Citation

  • Stiglitz, Joseph E.; and Andrew Weiss, 1981, “Credit Rationing in Markets with Imperfect Information,” The American Economic Review, Vol. 71, No. 3, June, pp. 393–410.

    • Search Google Scholar
    • Export Citation

  • Tanzi, Vito; and Hamid R. Davoodi, 2001, “Corruption, Growth, and Public Finances,” Political Economy of Corruption, edited by Arvind K. Jain, Routledge, London. pp. 89–110. Reprinted as chapter 11 in Governance, Corruption and Economic Performance, 2002, edited by George T. Abed and S. Gupta, International Monetary Fund, Washington DC.

    • Search Google Scholar
    • Export Citation

  • Tocqueville, Alexis de, 1835, Democracy in America, University of Chicago Press, 2000 (originally published in 1835).

  • UN, 2004A, United Nations Handbook on Practical Anti-Corruption Measures for Prosecutors and Investigators, Vienna, September.

  • UN, 2004B, United Nations Convention Against Corruption, United Nations Office on Drugs and Crime, New York.

  • Weill, Laurent, 2011, “Does corruption hamper bank lending? Macro and micro evidence,” Empirical Economics Vol. 41, Issue 1, Springer-Verlag, pp. 25–42.

    • Search Google Scholar
    • Export Citation

  • Williams, Barry, 2007, “Factors Determining Net Interest Margins in Australia: Domestic and Foreign Banks,” Financial Markets, Institutions and Instruments, pp. 145–165.

    • Search Google Scholar
    • Export Citation

  • Williamson, Oliver E., 2005, “The Economics of Governance,” a Richard T. Ely Lecture, The American Economic Association’s Papers and Proceedings, Vol. 95, No. 2, pp. 1–18, May.

    • Search Google Scholar
    • Export Citation

  • World Bank, Doing Business Reports, various annual reports, World Bank, Washington DC.

  • World Economic Forum, Global Competitiveness Report, various annual reports. Data available on: https://www.weforum.org/

Appendix I. Countries Included in the Sample

Table 7.

Countries Included in the Sample

article image
Note: The panel is unbalanced across time. Other countries were not included primarily due to data limitations.

Appendix II. Data Description and Sources

Table 8.

Data Description and Sources

article image
article image

Appendix III. Correlation Coefficients

Table 9.

Correlation Coefficients of Main Variables

article image
Table 9 (continued):

Correlation Coefficients of Main Variables

article image

Appendix IV. Robustness Checks

Table 10.

Robustness Check of Other Governance Indicators

article image
Table 11.

Robustness Check for Endogeneity Using Instrument Variables

article image
Table 12.

Robustness Check for Cross-Sectional Dependence

article image
Table 13.

Robustness Check for Non-Linearity

article image
Table 14.

Robustness Check with Lagged Dependent Variable

article image
Table 15.

Robustness Check for Outliers

article image
*

We are grateful for comments and suggestions from Bernardin Akitoby, David Amaglobeli, Mike Andrews, Pasquale Di Benedetta, Raju Huidrom, Srobona Mitra, Carlos Mulas-Granados, Hashem Pesaran, Nicola Pierri, Tigran Poghosyan, Khaled Sakr, Liliana Schumacher, Emil Stavrev, Johannes Wiegand and participants in an EUR Financial Sector Working Group seminar. Cristina Batog and Vizhdan Boranova provided excellent research assistance. Any errors remain our own.

1

We follow Knight’s (1921) definition of risk and uncertainty. While the probabilities of risks are known and can largely be hedged, uncertainty relates to more rare and unexpected events that cannot be hedged.

2

It is outside the scope of this paper to investigate whether governance quality, in addition to financial intermediation costs, also affects lending maturities, collateral requirements, etc. See, for example, Bae and Goal (2009) for such an analysis. We do not investigate intra-country heterogeneity (e.g. the impact of governance quality depending on bank ownership), as we use country-level rather than bank-level data. See, for instance, Fungáčová and Poghosyan (2011) for such an analysis. A different approach would have been to gauge the impact of the quality of institutions and good governance on sovereign spreads. See, for example, Akitoby and Stratmann (2010) for such an analysis.

3

Good governance is akin to the concept of social cost (Coase, 1960), good institutions (North, 1981, 1989; Acemoğlu, 2005; Acemoğlu and Robinson, 2008, 2012); social capital (Coleman, 1988; Knack and Keefer, 1997); social infrastructure (Hall and Jones, 1999); and social capability (Putterman, 2013). It is underpinned by social norms, ethics, and morals. It reduces transaction costs, including information costs. It alleviates negative externalities, market- and public-sector failures, and is conducive to enhancing the number of state- contingent markets. The latter permits further division of labor as well as risk diversification. Ultimately, enlightened self-interest (as discussed by Tocqueville, 1835) induces trust, which bestows more predictable outcomes, when reacting to unexpected shocks. Fukuyama (1995, p. 27) notes that “The legal apparatus, serving as a substitute for trust, entails what economists call “transaction costs”. Or as Arrow (1974, p. 23) noted: “Trust is an important lubricant of a social system.” In short, good governance enhances both output and stability. (See Annex Tables 2.1.1–3 in IMF 2017 for an overview of the theoretical and empirical literature.)

4

Evidently, a fair and free media is also important to ensure informed decisions as well as to advance checks and balances in both the government and private sector. Houston et al. (2011), for example, found that state- ownership of the media and media concentration were associated with more corruption in bank lending.

5

There are various definitions of corruption (UN, 2004; OECD, 2013; IMF, 2018). For the purpose of this paper, we use the following definition: “abuse of position to solicit an illicit gain.”

6

There are other proxies to measure efficiency of banks, like non-interest costs to total income (often called the “efficiency ratio”) or relative to total assets. However, non-interest costs will have to be covered by fee income and the NIM for a bank to remain profitable.

7

Bank concentration and competitiveness remain widely debated. Claessens and Laeven (2004), for example, argue that having contestable systems with a constant threat of entry, including from foreign banks, is more important for competition than low concentration. Based on a cross-country study, they found low correlation between the Panzar-Rosse (PR) index for bank competition and traditional measures, like the Herfindahl-Hirschman index and share of the largest banks.

8

However, more predictable enforcement could also imply better information about risky borrowers’ track record. In principle, this would reduce the credit rationing caused by imperfect and asymmetric information (Stiglitz and Weiss, 1981). In our view, the effect could potentially be mitigated by better screening and monitoring, which does not seem to be fully captured by the model.

9

The quality of banking supervision deserves special attention, but it is outside the scope of this paper. It is covered in other papers such as Beck, Demirgüç-Kunt and Levine (2006), Chortareas et al. (2012), and Barth et al. (2013).

10

See, among others, Demirgüç-Kunt and Huizinga (1999), although they find that the effect is weaker in wealthier countries, and Poghosyan (2013).

11

See, for instance Beck et al. (2006) and Akins et al. (2017).

12

Barth et al (2009) argue that better information, for instance via credit bureaus, leave less discretion for the loan officer and thus less room for bribery, as it increases the likelihood of exposure. Moreover, it reduces the rent banks can extract (Padilla and Pagino, 1997). Islam (2007) also finds that better information is positively related to higher levels of private credit, raises accountability and reduces corrupt bank lending. Akins et al. (2017) find that timely loan-loss recognition is associated with less lending corruption at the bank level, likely because timely recognition will expose poor credit policies at an earlier stage.

13

About 55 percent of potential observations are available for the specifications with the broadest coverage.

14

Governance indicators are often criticized for being based on perceptions and small samples. First, perceptions do matter, say when it is decided where to place a greenfield investment. Second, the proxies we use are composites from several sources. Third, so-called objective indicators must also be caveated. For instance, convictions for corruption may increase due to increased corruption or due to an intensified anti- corruption campaign. Note that the 2017 Special Eurobarometer on corruption found that on average 68 percent of the respondents in the 28 European Union (EU) countries believed that corruption was “very” or “fairly” widespread,” while only 12 percent knew somebody who had accepted or taken a bribe, and only 5 percent had experienced or witnessed corruption the last 12 months. Finally, despite such unease, it is noteworthy that governance indicators are generally highly correlated. For an overview of various governance indicators, see e.g., the Varieties of Democracy Institute.

15

Staikouras and Wood (2004) find for 685 European banks that banks transforming from small to medium-size typically become more efficient, while transforming from medium to large-size they often become less efficient. Claeyes and Vennet (2008), who also use the share of assets of the three largest banks to measure concentration, find that higher concentration is associated with lower NIM in non-EU accession countries, while the opposite was the case for banks in Western Europe and EU accession countries. Moreover, they note that the foreign bank entry to Central and Eastern Europe both led to bigger concentration and lower margins. Kasman et al. (2011) find that when controlling for direct efficiency measures, the coefficients for market share and concentration became insignificant for European banks during the 1995–2006 period.

16

The credit risk proxy only captures ex post credit risk, while the proxy for risk aversion is intended to measure buffers to cover unexpected, i.e. ex ante, losses stemming from credit, market and operational risks.

17

Claeyes and Vennet (2008), for instance, find a positive significant relationship for Western Europe, while the correlation is negative, although insignificant, for Central Eastern European banks. Liebeg and Schwaiger (2009) find—based on 895 local Austrian banks between the 1997–2005 period—that the NIM is mainly driven by bank characteristics rather than by macro variables, like changes in GDP and interest rate volatility.

18

The negative counter effects postulated by Jappelli et al. (2002 and 2005), thus seem muted.

19

The various governance indicators capture different aspects. For instance, the rule of law covers a broad spectrum, thus overlapping with other indicators, while corruption likely captures the impact of poor governance. Furthermore and related, the governance indicators we use are highly correlated, although they focus on particular aspects. Hence, if all governance indicators are included, we are likely to get multicollinearity.

20

Poghosyan (2013) looks at domestic bank competition and find a positive association with fraction of bank entries denied and higher margins for low-income and emerging market economies.

21

The perfect substitutability between debt and equity finance proposed by the Modigliani-Miller theorem may become complementary due to monitoring costs, etc. as argued by Boyd and Smith (1996). Obviously, the Modigliani-Miller and the other irrelevance theorems only work in a neoclassical phantom world with a complete set of state-contingent Arrow-Debreu markets, nicely behaved utility and production functions, etc. As soon as these assumptions become more realistic, asymmetric information, transaction and agency costs emerge and the quality of governance (Williamson, 2005) become relevant, while the irrelevance theorems become irrelevant.

22

They focused on interest spreads of OECD countries during the 1990–2015 period.

23

Alternative dependent variables, like interest rate spreads and ROA before taxation were also tried, but the explanatory power of the panel regressions is much weaker. Taxation could be an important missing explanatory variable. This is most likely to be the case for ROA, where taxation can be assumed to constitute a relatively larger share. Quantifying the impact of taxation is challenging due to lack of comparable data. The total tax rate on commercial profits reported by the Doing Business Reports was used as a rough proxy. However, some countries have special bank taxes, different rules for tax deductibility of provisions and deferment of losses, etc. that are not captured by this proxy. Moreover, the number of observations is limited. Furthermore, this rate is over 100 percent in some countries—typically countries in the lower end of the distribution of various governance indicators. All these factors may help explain the inconclusive results.

24

Admittedly, some bank characteristics and governance proxies are potentially correlated, like credit risk and the effectiveness of debt recovery. However, neither the variance inflation factor approach available in STATA nor correlation coefficients of the potentially correlated candidates indicated issues with multicollinearity (Appendix III).

25

This is extended by calculating bias-corrected least squares dummy variable (LSDV) estimators for the standard autoregressive panel data model using the bias approximations in Bruno (2005), who extends the results by Bun and Kiviet (2003), Kiviet (1999) and Kiviet (1995) to unbalanced panels. The results are obtained through bootstrapping variance-covariance matrix for LSDV estimators.

26

Additional robustness checks were performed. First, to examine if changes in governance are associated with changes in financial intermediation costs, we estimated specifications with some governance indicators in differences. The results confirm that better governance is typically associated with lower financial intermediation costs. Second, to examine if market structure is part of the underlying mechanism where governance shapes market structure, we estimated specifications excluding concentration and including some governance indicators. The results confirm that better governance is typically associated with lower financial intermediation costs. Third, to examine the robustness of results across subsamples, we followed Acemoğlu et al. (2001) and Demirguc-Kunt et al. (2004) and estimated specifications accounting for income groups. The results are available from the authors upon request.

27

Ideally, the different models should be estimated for each income group. Due to data availability, this was only possible for high-income countries, which reduced the annual savings for high-income countries.

28

Bae and Goyal (2009) reconfirmed that poor property rights and poor enforcement of these rights typically require more collateral, shorter maturities, and larger interest spreads. They used a sample of 63,158 loans of about 22,000 borrowers and banks in 48 countries during the period 1994–2003. They also found that lenders reduced maturities and increased lending spreads during the Asian Economic Crisis 1997–1998.

29

As Irving Fisher (1933, page 341) noted: “…over-confidence seldom does any great harm except when, as, and if, it beguiles its victims into debt.”

Same Series

Other IMF Content

Other Publishers

  • Collapse
  • Expand
Cover IMF Working Papers
Can Good Governance Lower Financial Intermediation Costs?
Author:
Mariusz Jarmuzek
and
Mr. Tonny Lybek
Volume/Issue:
Volume 2018: Issue 279
Publisher:
International Monetary Fund
ISBN:
9781484385678
ISSN:
1018-5941
Pages:
43
DOI:
https://doi.org/10.5089/9781484385678.001
Figure 1.

Net Interest Margins in Different Country Groups, 1996 – 2015
Figure 2.

Net Interest Margins over Time, 1996 – 2015